This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

lm3409 internal reference voltage

Other Parts Discussed in Thread: LM3409

Hi,

I want to build a LED constant current regulator to drive 7 LEDs, with a combined forward voltage 16.1V (OSRAM CP7P, @600mA). I tried to set the toff time to 295ns (Roff = 7.5k; Coff=470p), (I would like to modulate the EN pin ~ 2kHz)

However, the circuit produces a much higher off time, and noise peaks arise, while the IC is off. (the same noise is visible at Vout, PGATE...)

According the datasheet, the voltage at the Coff pin should be trimmed @1.24V. I measured the voltage there (10x probe, PicoScope 5000 series). As the image shows, it goes up to 7V (hence toff >2us)


 

Are there any external components (passive, power supply,...), which influence the voltage limit? What part of the circuit could be responsible for the noise peaks?

I appreciate any help, as this is my first experience with LED drivers.

  • Is there any way you could provide a schematic, BOM, and layout you are using? Is this the EVM with just the Roff resistor changed? What input voltage are you using?

  • By the way, you can send them (or whatever you are willing to share) in a private message if you do not want them public. But something is seriously wrong if the Coff pin is going that high. That is past the abs max and could damage the device.

  • Thank you for your response!

    I summarised the operation values:

    OPERATION POINT

    • Vin = 22.2V
    • Vout=16.1V
      • 7 Osram LH-Cp7P LEDs in series, Iled = 600mA

    VALUES:

    • Roff = 7.5k
    • Coff = 470p
    • Rsns=0.3 Ohm
    • Cf = 1uF
    • Cin = 100uF
    • L = 22uH
    • Ruv1 = 4.3k
    • Ruv2=51k
    • Diode: 60V, 3A
    • PFET: 30V, 3A

    The circuit schematics is as follows (+the EN pin was connected to Vini)

  • I don't see anything wrong with the design itself although I don't have the BOM to be sure of every component. So the things I would make sure of in that sense are that the inductor saturation current rating is higher than the peak inductor current and that the LM3409 is capable of driving the FET you are using at 1MHz which looks like the target frequency. The VCC current limit can be as low as 30mA, so make sure Qg*Fsw is less than that value. I also noticed the input capacitor is 100uF so I'm guessing it is aluminum electrolytic? If so I would at least put some ceramic in parallel to reduce the effective ESR (AEs have pretty high ESR so they do not filter noise on the input nearly as well). But most often if you see behavior that abnormal it is due to layout. So I have a few questions:

    1. Can you provide the BOM?

    2. I noticed you are using a DIP package. Is this circuit breadboarded or is it on an actual PCB?

    3. If it is on a PCB can you provide the layout for me to look over?

    More than half of the issues I see with switching regulators is due to layout. They generate significant noise so layout is very important. That is why I asked about breadboarding. A switching regulator will never work correctly if breadboarded.

  • BOM link
    Rsns ISHAY  LVR05R3000FE73  Strommesswiderstand, Durchsteckmontage, 0.3 ohm, Axial bedrahtet, 5 W, ± 1%, Baureihe LVR05 de.farnell.com/.../1108094
    Roff Nova resistor KIT, CBR14 uk.farnell.com/.../1198152
    Coff Noca capacitance KIT, CCC31,32 uk.farnell.com/.../1198143
    Ruv1
    Ruv2
    diode MULTICOMP  1N5821  SCHOTTKY DIODE, 3A, 30V, DO-27 de.farnell.com/.../1843516
    L wurth electronics Kit
    PFET INFINEON  BSS308PE H6327  MOSFET-Transistor, p-Kanal, -2 A, -30 V, 0.062 ohm, -10 V, -1.5 V de.farnell.com/.../2432719
    LED OSRAM LH-CP7P
    Cf
    Cin

    all the resistors capacitors are part Nova Kit


    2. I have heard, that a breadboard could cause problems in case of buck converters (circuits with high switching frequencies), hence I soldered the circuit on a test board (hoping, it helps). The signal forms remained the same.

    I did a PCB layout as well, but I am waiting with it, until I figure out, whats wrong, or how I could solve it

    I continued the measurements, and I just got more questions. (green- Uvcc, red-Uout, blue-Upgate; Picoscope 5000 series)

    As the block diagram indicates, PGATE is switched between Ucc and Uin (which corresponds to my measurenemts), and therefore Upgate amplitude is ~6V.  Some results from a test circuit are available online, where Upgate is switched from Uin to GND. (here)

    How does the regulator for pin Vcc work, and which voltage level should the pin reach?

  • Ok, I'm not 100% sure of the waveform I am seeing. Is this during PWM dimming? The most useful information would be the switch node waveform (Q, D, L intersection) and the LED current. The waveform above makes it hard to distinguish the switching frequency and how the output current is reacting. Having said that I reiterate layout is crucial. If you want to send me a picture of your circuit I would be happy to provide an opinion. But it isn't just breadboarding that can cause issues. Grounding is key, a good ground plane and following the recommendations in the datasheet is critical. One main spot of concern is the grounding of Cin, the IC, and the rectifier diode. You are trying to switch at 1MHz which requires very careful layout. Of one more concern is that you are using a wirewound current sense resistor (I think). Those have excessive parasitic inductance and can never be used for a switching regulator working at over 10s of kHz or so maximum. Like I said, layout is a common issue. Even on a PCB it can be easy to inject enough noise to make the regulator misbehave (this includes all switching regulators from any company). If you would like me to check your PCB layout before fabrication and give recommendations I would be happy to. You really need to have a solid layout before you can diagnose any other potential problems and I think your circuit, values anyway, is just fine.

  • I checked the Cin-Rsns-PFET-diode loop (TURNED OUT TO BE EXTREMELY CRITICAL) on my board , which actually went around the lm3409(no wonder, it misbehaved; looks like I neglected that paragraph in the datasheet...). Therefore I re-designed my test board, keeping this loop as small as possible, and now the both Coff-pin, and the switch node voltage looks fine (cropped @1.24V, and square wave, respectively).

    However, I would be happy, if I could send you my layout in private, to make sure, it's all right.
  • Sure. I will accept your friend request so you can send private messages.